Journal of Personality and Social Psychology
2008, Vol. 95, No. 6, 1541–1557
Copyright 2008 by the American Psychological Association
0022-3514/08/$12.00 DOI: 10.1037/a0013334
Assertiveness Expectancies: How Hard People Push Depends on the
Consequences They Predict
Daniel R. Ames
Columbia University
The present article seeks to explain varying levels of assertiveness in interpersonal conflict and
negotiations with assertiveness expectancies, idiosyncratic predictions people make about the social and
instrumental consequences of assertive behavior. This account complements motivation-based models of
assertiveness and competitiveness, suggesting that individuals may possess the same social values (e.g.,
concern for relationships) but show dramatically different assertiveness due to different assumptions
about behavioral consequences. Results clarify the form of assertiveness expectancies, namely that most
people assume increasing assertiveness can yield positive social and instrumental benefits up to a point,
beyond which benefits decline. However, people vary in how assertive this perceived optimal point is.
These individual differences in expectancies are linked in 4 studies to assertiveness, including selfreported assertiveness, rated behavioral preferences in assorted interpersonal conflict scenarios, partner
ratings of participants’ behavior in a face-to-face dyadic negotiation, and work colleague ratings of
participants’ assertiveness in the workplace. In each case, the link between expectancies and behavior
remained after controlling for values. The results suggest a place for expectancies alongside values in
psychological models of interpersonal assertiveness.
Keywords: assertiveness, competitiveness, expectancies, conflict, interpersonal schemas
ation that complements a motivational account. I argue that a
person’s expectations about the consequences of assertiveness
predict their behavior in conflicts above and beyond the effects of
their values. In other words, sometimes tough people are tough
because they expect they can be, and soft people are soft because
they believe they must be. The results of several studies support
this account and hold implications for studying, counseling, and
teaching about the dynamics of conflict.
Why, in interpersonal conflicts, do some people come on strong,
make bold demands, and concede little ground, whereas others
give way, yielding to another person’s requests and surrendering
their own interests? Over the last generation, the literatures on
conflict, negotiation, and social dilemmas have highlighted the
importance of values in answering this question, tracing higher
levels of assertiveness to proself orientations and competitive
conflict styles and lower levels of assertiveness to prosocial orientations and cooperative conflict styles (e.g., Carnevale & De
Dreu, 2006; De Dreu, Weingart, & Kwon, 2000; Komorita &
Parks, 1996; Van Lange, 1999). The predominant answer has been,
in effect, that tough people are tough because they care more about
winning, whereas soft people are soft because they just want to get
along.
There is little doubt that values vary from person to person, and
such variance explains meaningful differences in behavior. Yet, as
learning theories dating back at least to Tolman (1932) have noted,
and more recent social cognitive models of behavior (e.g., Bandura, 1986; Mischel & Shoda, 1995) have recognized, people’s
behavior also reflects their expectations about the outcomes of
their acts. People base their action not only on what they want but
also on what they expect their behavior will achieve. The present
article draws on this long tradition of work to pursue an
expectancy-based account of assertiveness in conflict and negoti-
Values and Conflict
Over the last half century, scholars of conflict have repeatedly
linked interpersonal conflict behavior to underlying values—
variously identified as preferences, concerns, priorities, and orientations. One active tradition of work revolves around dual-concern
theory (e.g., Carnevale & Pruitt, 1992) and motivational orientations (e.g., Messick & McClintock, 1968), which posit that people
vary in their attitudes about their own and their conflict partners’
outcomes. Combinations of these dimensions yield different orientations such as “proself” (maximizing the difference between
self and other) and “prosocial” (maximizing joint outcomes). Considerable research has linked these social value orientations to
assertive behaviors in social dilemmas (e.g., McClintock &
Liebrand, 1988; Van Lange, 1999) and in conflict and negotiation
(e.g., De Dreu & Van Lange, 1995; De Dreu et al., 2000; Olekalns
& Smith, 2003).
Elsewhere, work in the interpersonal circumplex tradition has
linked differences in concerns with agency and communion to
assertiveness in conflict and negotiation (e.g., Amanatullah, Morris, & Curhan, in press; Locke & Sadler, 2007). Scholars have also
invoked individual differences in need to belong (e.g., De Cremer
& Leonardelli, 2003) and face-saving concerns (e.g., White,
I thank Emily Amanatullah, Emily Bianchi, Jeremy Burrus, Adam
Galinsky, Lara Kammrath, Elke Weber, and, especially, Brian Lickel for
thoughtful comments on the research and manuscript.
Correspondence concerning this article should be addressed to Daniel
Ames, Columbia Business School, Columbia University, 707 Uris Hall,
3022 Broadway, New York, NY 10027. E-mail: [email protected]
1541
AMES
1542
Tynan, Galinsky, & Thompson, 2004) to explain conflict behavior.
In short, ample evidence shows that what people care about affects
their assertiveness. There is no reason to question the usefulness of
individual differences in values in accounting for interpersonal
assertiveness, but there are reasons to suspect they are not the
whole story. People’s ability to explain differences in assertiveness
could be enhanced by incorporating another construct psychologists have long recognized in other domains (e.g., Bandura, 1977):
outcome expectancies.
Outcome Expectancies and Assertiveness
Although expectancies take many forms—ranging from stereotypes to attachment models (see Roese & Sherman, 2007, for a
review)—the present account stresses expected reactions by another to one’s own assertive behavior. “If I do X,” an actor thinks
to her- or himself, “then this other person will do, think, or feel Y.”
Building on social cognitive models of behavior (e.g., Bandura,
1986; Mischel & Shoda, 1995), such outcome expectancies have
been examined in several contemporary traditions of work. Research on relational schemas has shown that behavior, such as the
amount of anger displayed to a partner, depends in part on anticipated partner response (Fehr, Baldwin, Collins, Patterson, &
Benditt, 1999; see also Baldwin & Dandeneau, 2005, for a recent
review). A related area of inquiry has examined rejection sensitivity, which revolves around “anxious expectations” of interpersonal rejection and the associated activation of defensive responses, which can have negative or even self-fulfilling effects
(e.g., Downey & Feldman, 1996).
These and other areas of research suggest that expectancies will
influence assertive behavior in interpersonal conflicts. Yet, the
research to date has largely focused, for good reasons, on expectancies revolving around specific others, such as romantic partners
(e.g., Kammrath & Dweck, 2006), or has been based on attributes
such as race or sex (e.g., Amanatullah, 2008; Bowles, Babcock, &
Lai, 2007). As such, the existence and impact of general assertiveness expectancies remains largely unexplored by expectancy researchers as well as by scholars of negotiation and conflict. The
studies presented here pursue this topic directly.
The Nature of Assertiveness Expectancies
In the present account, I attempt to explain differing individual
levels of assertiveness in everyday interpersonal conflict behaviors, including dyadic negotiations. Assertiveness is taken here to
mean a person’s tendency to stand up and speak out for their own
interests and concerns, such as voicing opinions, making offers
and concessions, and attempting to coerce or intimidate others (cf.
Ames & Flynn, 2007). To predict such behaviors with assertiveness expectancies, the nature of these beliefs (e.g., what is the
relevant expectation to assess?) must first be clarified.
On the basis of prior work showing that interpersonal assertiveness often has a curvilinear, inverted-U-shaped effect on interpersonal relations (Ames & Flynn, 2007), I expect that the majority of
people will have curvilinear expectancies, assuming that they can
push up to a point, but no further without incurring damage to their
outcomes and/or relationships. For instance, in a negotiation, people may feel that making a moderately assertive opening offering
could be effective, but at some point of heightened assertiveness,
an opening could backfire, undermining both results and relationships. Although people in general may show this form of expectancy, individuals vary in what point they think they can push up to.
Some may be very optimistic, assuming they can display high
levels of interpersonal assertiveness before incurring costs. Others
may be much more pessimistic, assuming that even modest levels
of assertiveness could spell trouble. If this characterization is
correct, then it would invite a research approach that gauges what
individuals expect would happen when they push hard or give in.
An important initial goal of the present work, then, was to
establish the nature of assertiveness expectancies. I suspect a
majority of people will show a curvilinear pattern, believing that
increasing assertiveness brings benefits up to a point. Just as
important, though, I expect parameters of these expectancies (in
particular, the assumed optimal level of assertiveness) to vary from
person to person and that this variance can be of use in predicting
behavior. An exploratory concern in the present article is how to
measure these individual differences in assertiveness expectancies
practically and effectively.
Assertiveness Expectancies and Behavior
The present account offers a central prediction: People make
idiosyncratic forecasts of the social and instrumental consequences
of a given forceful or acquiescent behavior, and, as social cognitive theories of personality suggest (Bandura, 1977; Mischel,
1973), these outcome expectancies shape behavior and can help
researchers to predict it. It is possible that assertiveness is predominantly the product of values and that expectancies are simply
reflections of these values. That is, those who expect high assertiveness to be harmful could simply be more concerned with
preserving relationships. On the contrary, I expect that whereas
assertiveness expectancies might be related to values, an independent expectancy– behavior link will remain after controlling for
values. Two people could have identical values but differ in
assertiveness simply because one expects a behavior would damage a relationship and the other does not.
The present account does not imply that values are unimportant,
only that values will not account for all of the effects of expectancies. It is also worth noting that expectancies and values may
interact: The expectancies that matter most in predicting behavior
may be the ones concerning highly valued outcomes. Such interactions are examined in the analyses presented here.
Plan of Study
This article proceeds in two cumulative stages. The first stage is
partly exploratory, examining how assertiveness expectancies
might best be characterized and measured. The second stage is
confirmatory, basing its operationalization of expectancies on the
results of the first stage and attempting to explain assertive behavior with individual expectancies. Studies 1 and 2 comprise the first
stage, inviting participants to literally chart out their expectancies
and to consider a series of responses to hypothetical conflicts.
Studies 3 and 4 comprise the second stage, using expectancies to
predict behavior and outcomes in a dyadic negotiation and in the
workplace.
ASSERTIVENESS EXPECTANCIES
Study 1
In Study 1, I examined the nature of assertiveness expectancies
by asking participants to draw lines showing social and instrumental outcomes as functions of a wide spectrum of assertiveness. The
resulting figures were categorized by shape and analyzed for
properties such as perceived optimal assertiveness. Measures of
expectancies were compared with social values and self-reported
assertiveness. I expected that the majority of lines would show
inverted U curvilinear shapes and that individual differences in
these expectancies (e.g., perceived optimal assertiveness) would
predict self-reported assertiveness after controlling for values.
Method
Participants. To assess the generalizability of results, Study 1
drew on two populations: paid university research participants and
master’s of business administration (MBA) students. No differences were expected between the groups. A total of 167 participants completed Study 1, including 104 MBA students, who took
part in a course exercise involving the materials from Study 1, and
63 members of an East Coast university paid research participation
program, the vast majority of whom were undergraduate students.
Mean age for the MBA sample was 29.0 (SD ⫽ 3.0); 72 (69.2%)
were men. Of the 93 MBAs who indicated ethnicity, 52 (55.9%)
identified themselves as Caucasian, 29 (31.1%) identified themselves as Asian or Asian American, 3 (3.2%) identified themselves
as Latino or Hispanic, and 3 (3.2%) identified themselves as
African American. Mean age for the paid participants was 21.6
(SD ⫽ 4.2), and 38 (60.3%) identified themselves as women. Of
the 56 who identified their ethnicity, 22 (39.3%) identified
themselves as Caucasian, 19 (33.9%) identified themselves as
Asian or Asian American, 8 (14.3%) identified themselves as
African American, 6 (10.7%) identified themselves as Latino or
Hispanic, and 1 (1.8%) identified him- or herself as indigenous/
Native American.
Materials. Participants recorded their expectancies by drawing lines or curves related to two domains: social outcomes (e.g.,
being liked and trusted by others) and instrumental outcomes (e.g.,
getting favorable settlements and deal terms). Recent work by
Ames and Flynn (2007) suggests that actual outcomes in these
domains sometimes diverge, such that increasing assertiveness can
sometimes yield instrumental gains and social costs. However, it
was not clear, a priori, whether expectancies in one or the other of
these domains would be a superior predictor of assertive behavior.
As such, both were measured, and an exploratory consideration
throughout the present article was whether one construct or the
other better predicted behavior.
Participants were presented with hard-copy survey materials
featuring x- and y-axes forming the bottom and left sides of a chart.
The bottom of the chart was labeled Assertiveness and ranged from
0 to 100 in increments of 10. For the social outcomes chart, the
y-axis was labeled social outcomes (trusted/liked by partner, partner wants to interact again in future). Participants were told that in
this case, “partner” referred to the other (opposing) party in a
conflict or negotiation. Instructions read as follows:
How do you think a person’s assertiveness is related to their social
outcomes in a conflict or negotiation, such as a negotiation partner’s
liking and trust for them, and their partner’s interest in interacting with
1543
them again in the future? In the blank graph below, draw a line that
represents how you see this relationship: what is the effect of assertiveness on social outcomes? The line can be any shape or angle you
want, straight or curved. Your line should span the whole width of the
chart, from the 0th percentile in assertiveness to the 100th percentile.
For the instrumental outcomes chart, the y-axis was labeled
Results ( payoffs, deal terms, and getting one’s way). Instructions
read as follows:
How do you think a person’s assertiveness is related to their results in
a conflict or negotiation, non-social outcomes including their payoffs,
deal terms, and the extent to which they get their way? In the blank
graph below, draw a line that represents how you see this relationship:
what is the effect of assertiveness on results? The line can be any
shape or angle you want, straight or curved. Your line should span the
whole width of the chart, from the 0th percentile in assertiveness to
the 100th percentile.
The order of charts was counterbalanced. Pilot work revealed
that participants found the task meaningful and captured assumptions that they “knew” but otherwise found difficult to articulate.
Pilot work using z scores for the axes led to confusion; participants
found the task most interpretable and meaningful when the axes
were described in terms of percentiles for their peer group.
Along with the line drawing, participants indicated their own
assertiveness as well as their own social values. For self-rated
assertiveness, participants completed a percentile measure:
Compared to your classmates, what percentile do you think you’re in
for assertiveness? 50% would mean average assertiveness compared
to classmates, while 10% would mean you’re in the lowest 10%, and
90% would mean you’re in the top 10% in terms of assertiveness.
For social outcome values, participants were asked, “Sometimes,
people care more about their results or more about their social
outcomes. How would you describe your general balance of these
preferences? Do you generally care about results or outcomes
more— or care about them the same?” Participants rated their
response on a 13-point scale ranging from 1 (Much more about
results) to 13 (Much more about social outcomes), with the midpoint, 7, labeled About the same. The order of the self-reported
assertiveness and value questions was counterbalanced, as was
their placement before versus after the line drawing exercise.
Results
Results for the two participant populations were very similar. As
such, the samples were collapsed for analyses reported below
except where noted.
Characterizing expectancies. Study 1 was concerned with
characterizing assertiveness expectancies with an eye toward identifying effective ways to measure them. As anticipated, many
participants expected curvilinear effects of assertiveness on both
social and instrumental outcomes such that both extremely low and
high levels of assertiveness were associated with lower outcomes.
The majority of both paid and MBA participants drew invertedU-shaped functions for social and instrumental outcomes (see
Table 1). The mean anticipated outcomes across all participants are
shown in Figure 1. It is worth noting that although most participants indicated an inverted U curve, a sizable minority (33%) drew
AMES
1544
Table 1
Frequency of Shapes of Individual Curves Depicting Outcomes as a Result of Assertiveness In Study 1
Social outcomes
Instrumental outcomes
Variable
Paid
participants
MBA
participants
All
participants
Paid
participants
MBA
participants
All
participants
Inverted U
Downward convex
Downward concave
Downward straight
Upward convex
Upward concave
Upward straight
Other
77.8
1.6
1.6
3.2
7.9
3.2
3.2
1.6
63.5
2.9
10.6
5.8
7.7
3.8
2.9
1.9
68.9
2.4
7.2
4.8
7.8
3.6
3.0
1.8
54.0
0.0
0.0
1.6
20.6
7.9
15.9
0.0
66.3
1.0
1.0
1.0
12.5
6.7
6.7
4.8
61.7
0.6
0.6
1.2
15.6
7.2
10.2
3.0
Note. Cells represent percentage of participants (paid participant n ⫽ 63, MBA participant n ⫽ 104). Inverted U functions showed clear downturns at
both the low and high ends; convex lines showed an upward bulge; concave lines showed a downward bulge; straight lines showed no curvature. MBA ⫽
Master’s of business administration.
upward lines with no discernable downturn for instrumental outcomes as a function of assertiveness (see Table 1).
The prevalence of curvilinear expectancies suggests that one
way to capture individual differences in assertiveness expectancies
is to gauge individuals’ perceived optimal level of assertiveness,
the point of assertiveness that a person believes maximizes outcomes (i.e., pushing up to this point brings net benefits, but
pushing beyond it undermines results). For a linear upward slope
for instrumental outcomes, this “instrumental optimum” was 100.
For an inverted U line in which social outcomes were greatest at an
assertiveness level of 60, this “social optimum” was 60. Higher
optima thus indicated expectancies that higher levels of assertiveness would afford better outcomes. Put another way, perceived
optimal assertiveness reflects the point at which people believe
they can gainfully push up to. Analyses reported below test
whether these optimum measures predicted self-reported assertive-
IInstrumental
t
t l
outcomes
80
Expe
ected outtcomes
70
60
50
40
30
Social
outcomes
20
10
0
0
10
20
30
40
50
60
70
80
90
100
ness, but before attempting to predict assertiveness, additional
explorations of assertiveness expectancies as a construct and measure are worth considering.
Individual differences in expectancies are necessary to explain
individual differences in behavior. Thus, a natural question is: Are
there areas in which the variance in expectancies is more or less
pronounced? To gauge this, I plotted the mean level of anticipated
social and instrumental outcomes across the spectrum of assertiveness, as well as values one standard deviation above and below the
mean, in the top half of Figure 2. As can be seen in the top left
portion of the figure, social expectancies seem to have somewhat
greater variance at the extremes of assertiveness. In the top right
portion of the figure, the variance of instrumental expectancies
seems clearly higher at the very highest levels of assertiveness. In
other words, people appeared to agree that low assertiveness leads
to poor instrumental outcomes. However, people diverged about
the instrumental rewards they anticipated for the highest levels of
assertiveness.
Expectancy constructs. On the basis of the foregoing analyses
attempting to characterize assertiveness expectancies, three types
of measures were computed as independent variables that might
predict self-reported assertiveness. The first of these was perceived
optimal assertiveness, as described above (i.e., the perceived level
of assertiveness that maximized social or instrumental outcomes).
The other two measures of expectancies were based on expected
outcomes for extreme low and high levels of assertiveness. Specifically, anticipated outcomes (as reflected in participants’ line
drawings) for the 0 and 10 levels of assertiveness were averaged to
compute measures of expectancies for low assertiveness. Likewise, anticipated outcomes for the 90 and 100 levels of assertiveness were averaged to compute measures of expectancies for high
assertiveness.1 Separate measures were computed for social and
instrumental outcomes.
Predicting assertiveness. The constructs described above were
analyzed for their ability to predict self-reported assertiveness.
This affords a test of the prediction that expectancies predict
Assertiveness
Figure 1. Mean ratings for expected social and instrumental outcomes
across levels of assertiveness in Study 1.
1
Results were very similar for just the extreme values, 0 and 100, as
well as an average of three endpoints (0, 10, and 20; 80, 90, and 100).
ASSERTIVENESS EXPECTANCIES
1545
Where do expectancies vary most from person to person?
Plots of means and one standard deviation above/below mean
Social outcomes
80
+1 SD
60
Mean
40
20
-1
1 SD
0
0
10
20
30
40
50
60
70
Instrumental outcomes
100
80
90
100
-20
Expec
cted instrumental outcomes
Expected social outcomes
s
100
80
+1 SD
Mean
60
40
20
-1 SD
0
0
10
20
30
40
50
60
70
80
90
100
90
100
-20
Assertiveness
Assertiveness
Which expectancies best predict self-reported assertiveness?
Across participant correlation between self-reported assertiveness and
anticipated outcomes for each level of assertiveness
Social outcomes
0.2
0.1
0
0
10
20
30
40
50
60
70
Instrumental outcomes
0.3
80
90
100
-0.1
-0.2
-0.3
Correlation with self-reported
assertiveness
s measure
Correlation with self-reported
s measure
assertiveness
0.3
0.2
0.1
0
0
10
20
30
40
50
60
70
80
-0.1
-0.2
-0.3
Assertiveness
Figure 2.
Assertiveness
Variance in expectancies and correlation with optimum measures in Study 1.
assertiveness and also informs the exploratory effort to identify
what measures of expectancies are the most effective predictors.
As shown in Table 2, the social optimum measure (perceived
optimal assertiveness for social outcomes) was positively corre-
lated with self-rated assertiveness, meaning that individuals who
expected they could push harder before incurring social costs were
also more likely to self-report higher levels of assertiveness. In addition to social optimum, both low- and high-assertiveness social
AMES
1546
Table 2
Correlations Between Curve-Based Constructs, Values, and Self-Rated Assertiveness in Study 1
Variable
1
1. Social optimum
2. Social expectancies for
low assertiveness
3. Social expectancies for
high assertiveness
4. Instrumental optimum
5. Instrumental expectancies
for low assertiveness
6. Instrumental expectancies
for high assertiveness
7. Social outcome values
8. Self-rated assertiveness
—
M
SD
Note. n ⫽ 162.
p ⬍ .10. ⴱ p ⬍ .05.
†
ⴱⴱ
2
⫺.80ⴱⴱ
.68ⴱⴱ
.22ⴱⴱ
⫺.12
ⴱ
3
4
—
.24ⴱⴱ
—
5
6
7
8
—
⫺.42ⴱⴱ
⫺.15ⴱ
.26ⴱⴱ
⫺.02
ⴱⴱ
⫺.48ⴱⴱ
ⴱⴱ
—
.19
⫺.03
.25ⴱⴱ
⫺.08
⫺.01
⫺.19ⴱ
.40
⫺.01
.23ⴱⴱ
.65
⫺.04
.14†
⫺.16ⴱ
.00
⫺.09
—
⫺.03
.06
—
⫺.25ⴱⴱ
—
54.12
29.79
26.62
28.40
36.65
29.47
78.74
21.59
11.27
14.44
64.27
28.88
6.15
2.42
60.71
18.11
p ⬍ .01.
expectancies were predictive. Those who believed very low levels
of assertiveness had more positive social outcomes reported lower
levels of personal assertiveness. Those who believed very high
levels of assertiveness had more positive social outcomes reported
higher levels of personal assertiveness. It is worth noting that when
both low- and high-assertiveness social expectancies were entered
as independent variables in a regression predicting self-reported
assertiveness, high assertiveness expectancies remained significantly predictive (␤ ⫽ .18), t(163) ⫽ 2.15, p ⬍ .05, but low
assertiveness expectancies appeared to lose their predictive power
(␤ ⫽ ⫺.11), t(163) ⫽ 1.36, p ⫽ .18. In short, several operationalizations of social expectancies were predictive of self-reported
assertiveness (see the bottom half of Figure 2 for additional detail
on how expectancies at different levels of assertiveness predicted
self-reported assertiveness). As shown in Table 2, the instrumental
expectancy measures were not significantly correlated with selfreported assertiveness.
Independent effect prediction. With the social outcome value
measure, Study 1 also afforded a test of whether the expectancy–
assertiveness link remained after controlling for values. As might
be expected, the value measure (relative concern with social outcomes) was negatively correlated with self-rated assertiveness,
though this measure was not significantly correlated with the
expectancy constructs (see Table 2). To test the independent effect
prediction (that the predictive power of expectancies would remain
after controlling for values), regression analyses were conducted,
predicting self-reported assertiveness with social expectancies and
values simultaneously. In a first model, both social optimum (␤ ⫽
.25), t(160) ⫽ 3.41, p ⬍ .01, and values (␤ ⫽ ⫺.22), t(160) ⫽
⫺2.98, p ⬍ .01, were predictive. In a second model, both social
expectancies for low assertiveness (␤ ⫽ ⫺.20), t(160) ⫽ ⫺2.68,
p ⬍ .01, and values (␤ ⫽ ⫺.24), t(160) ⫽ ⫺3.22, p ⬍ .01, were
predictive. In a third model, both social expectancies for high
assertiveness (␤ ⫽ .23), t(160) ⫽ 3.11, p ⬍ .01, and values (␤ ⫽
⫺.24), t(160) ⫽ ⫺3.19, p ⬍ .01, were predictive. These results are
consistent with the independent effect prediction.
Additional models were run to test for the possible interaction
between expectancies and values (i.e., that social outcome values
affect whether social or instrumental expectancies will be more
predictive of self-rated assertiveness). A wide range of models
revealed no significant interaction effects.
Discussion
A central goal of Study 1 was to characterize peoples’ expectancies about interpersonal assertiveness, attempting to shed light
on how they might be conceptualized and measured effectively in
the service of predicting individual differences in assertive behavior. As anticipated, and echoing what Ames and Flynn (2007)
found in their work on the actual impact of assertiveness, the
majority of participants in Study 1 displayed curvilinear social and
instrumental expectancies, assuming they could push up to a point,
beyond which costs would begin to outweigh benefits. More
important, these expectancies varied from person to person. One
point of variance was the perceived optimum, the point to which
people believed they could push—that is, the level of assertiveness
that maximized outcomes. Analyses also revealed that people
tended to agree more on the expected outcomes for moderate
assertiveness, diverging more significantly in their forecasts of
what would follow from extreme low or high assertiveness.
Study 1 also afforded an initial test of whether expectancies predict
assertiveness. All three measures of social expectancies predicted
self-reported assertiveness, including social optimum (perceived optimal assertiveness for social outcomes), social expectancies for
low assertiveness, and social expectancies for high assertiveness.
These links remained after controlling for social outcome values
(i.e., concern with relationships vs. instrumental outcomes). Indeed, social expectancies were not significantly correlated with
outcome values, casting doubt on the notion that the two constructs
simply reflect the same underlying attitude.
Instrumental expectancies did not predict self-rated assertiveness in Study 1. It could be that instrumental expectancies simply
matter less to people than social ones. People may prioritize the
preservation of relationships and find perceived relational costs
relatively abhorrent (e.g., Baumeister & Leary, 1995). It could also
be, though, that the approach taken in Study 1’s line drawing
ASSERTIVENESS EXPECTANCIES
paradigm was vague in terms of what was meant by “results.”
Perhaps when perceivers consider a concrete situation with specific instrumental outcomes that they can identify (e.g., a starting
salary in a job negotiation, as in Study 2), the ability of instrumental expectancies to predict behavior will be greater.
Study 2
Study 1 shed light on how assertiveness expectancies could be
operationalized, and links between expectancies and self-reported
assertiveness were found. Although encouraging, the measures
used were rather decontextualized: The independent measures of
expectancies were based on line drawings about the general impact
of assertiveness, and the dependent measure was a self-report of
general assertiveness. Before committing to an operationalization
of expectancies on the basis of line drawings alone, Study 2 was
used to examine expectancies with greater specificity and contextual detail. Study 2’s more concrete approach measured expectancies and behavioral preferences in a series of specific scenarios,
such as a salary negotiation.
Study 2 had two goals. A first goal was to test the prediction that
expectancies would predict assertiveness and that this link will
remain after controlling for relevant values; accordingly, additional measures of relevant values were introduced. The second
goal was to extend the exploratory work from Study 1 on gauging
assertiveness expectancies. Recall that in Study 1, expected outcomes for low and high levels of assertiveness showed considerable variance and were variously linked to self-reported behavior.
Accordingly, the focus in Study 2 was on expected outcomes for
low and high levels of assertiveness as a way of gauging expectancies. In Study 2, I used expectancies in particular contexts to
predict behavioral preferences in those same contexts. Both social
and instrumental expectancies were measured, allowing an assessment of whether either or both were predictive.
Method
Participants. Sixty-seven working managers enrolled in a
business education program in the United States participated in
Study 2 as part of a session exercise. Mean age was 37.2 (SD ⫽
5.27), and 54 (80.6%) identified themselves as male. Forty-seven
(70.1%) identified themselves as Caucasian, 15 (22.4%) identified
themselves as Asian or Asian American, 1 (1.4%) identified himor herself as African American, 1 (1.4%) identified him- or herself
as Native American, and 3 (4.5%) did not identify their ethnicity.
Assertiveness scenarios. Scenarios were used to assess both
assertiveness expectancies (the independent measure) and behavioral preferences (the dependent measure). Participants responded
to three scenarios presented in counterbalanced order: one featuring a salary negotiation, one regarding asking a freelancer for help
under a deadline, and one concerning disagreeing with a colleague
in a team meeting. Each scenario featured a beginning (“Imagine
that you were negotiating with a potential employer . . . ”) and then
three endings featuring low, medium, and high assertive behaviors.
The scenarios and endings are presented in the Appendix. Endings
were presented in counterbalanced order.
After each ending, participants rated expected social outcomes
(“How much do you think the [manager, freelancer, work colleague] would like, trust, and want to interact with you in the
1547
future?”) on an 11-point scale ranging from 1 (strong dislike,
distrust, disinterest) to 11 (strong like, trust, and interest). Participants also rated expected instrumental outcomes (e.g., “How well
do you think you would do in terms of final negotiated salary”) on
an 11-point scale ranging from 1 (extremely poorly) to 11 (extremely well). These anticipated social and instrumental outcomes
at the scenario level were taken as the independent variable measures of expectancies. That is, expectancies for high assertiveness
were the ratings for social outcomes and instrumental outcomes for
the high assertiveness ending. As noted in the Results section,
separate multilevel models were run using low, medium, and high
assertiveness expectancies as predictors, as well as all of these
expectancies as simultaneous predictors.
After rating all three endings for a given scenario, participants
were asked to consider the endings and “rank these three responses
for how likely you would be to use them.” Participants were told
to give a 1 for the response they would be most likely to use and
a 3 for the response they would be least likely to use. These
rankings were used to construct a dependent measure of preferred
assertiveness at the scenario level. This was computed as the level
of assertiveness (low, moderate, or high) that a respondent most
preferred for a given scenario. Low assertiveness was assigned a
value of 1, moderate assertiveness a value of 2, and high assertiveness a value of 3. Thus, participants’ preferences for higher
assertive responses in the scenarios were reflected in higher scores
for preferred assertiveness. The average preferred assertiveness
across the scenarios (including 198 valid responses; three per
participant) was 2.36 (SD ⫽ 0.50). In the majority of cases
(61.6%), respondents indicated a preference for the moderate assertiveness response. In nearly all of the remaining cases (37.4%),
respondents indicated a preference for the high assertiveness response; respondents most preferred the low assertiveness response
in only 1% of cases.
Value measures. In a separate exercise completed earlier in the
business education program, participants responded to items gauging relevant values. Two items gauged social outcome values, a
comparative concern for relational outcomes over instrumental
ones, including “I care more about getting along with others than
‘winning’ each conflict or disagreement” and “In a conflict, I
generally care more about results (getting my way) than social
outcomes (my partner’s liking and trust for me, and interest in
interacting with me again in the future).” Participants rated these
items on 5-point scales ranging from 1 (disagree strongly) to 5
(agree strongly).
In addition, participants completed a measure of unmitigated
communion, a construct involving extreme concern for others
(Fritz & Helgeson, 1998) that has been linked to negotiation
outcomes (Amanatullah et al., in press). Because of time constraints, five items were sampled from the original nine-item scale
(“I always place the needs of others above my own,” “For me to
be happy, I need others to be happy,” “I can’t say no when
someone asks me for help,” “Even when exhausted, I will always
help other people,” “I often worry about others’ problems”). Items
were rated on 5-point scales ranging from 1 (disagree strongly) to
5 (agree strongly).
Finally, in a separate exercise, participants indicated their social
value orientations (SVO), a commonly used tool for measuring
values relevant to conflict and negotiation (e.g., Van Lange, 1999).
The SVO measure gathered responses to a series of nine decomposed
1548
AMES
games (Messick & McClintock, 1968; see also Van Lange, Otten,
De Bruin, & Joireman, 1997). In each game, participants chose
between three alternative pairs of outcomes, each representing one
of the three SVOs: prosocial (maximizing joint gain), individualist
(maximizing own gain without regard for the other’s outcome),
and competitive (maximizing the difference between own and
other outcome). The number of responses consistent with each
orientation was used to create continuous measures for each of the
three orientations.
Results
The primary objective in Study 2 was to use expectancy and
other measures to predict preferred assertiveness. Due to the nested
nature of the data, multilevel modeling was used. It was not
expected that scenario would matter to the expectancy–
assertiveness link, yet controlling for scenario could clarify the
results. A Level 1 model was created at the scenario level (three
per participant), featuring preferred assertiveness as the dependent
measure (i.e., the level of assertiveness ranked as most preferred,
where low assertiveness ⫽ 1, moderate assertiveness ⫽ 2, and
high assertiveness ⫽ 3). This was predicted by dummy codes for
scenario (two dummy codes to reflect the three scenarios) and
social and instrumental expectancies for the given scenario. Separate models featuring low, moderate, and high assertiveness expectancies were run, as described in the paragraph below. The
Level 2 model at the person level featured effects for the value
measures, including unmitigated communion, SVO competitive,
and SVO prosocial (given the fixed sum structure of the SVO
measure, the SVO individualist variable was withheld from the
model because it is equivalent to the residual of the SVO competitive and prosocial measures). In addition, because the social value
items (getting along with others, caring about results) were only
modestly correlated with one another, r(66) ⫽ .45, they were both
entered as independent predictors in the Level 2 model.
Using hierarchical linear modeling (HLM) 6 to compute least
squares estimates of fixed effects (Raudenbush, Bryk, Cheong, &
Congdon, 2004), a first model featured high-assertiveness social
and instrumental expectancies in Level 1 (the scenario level). In
this model, both social and instrumental expectancies were significant predictors (unstandardized coefficents of 0.047, t[189] ⫽
3.52, p ⬍ .01; and, 0.058, t[189] ⫽ 4.78, p ⬍ .01, respectively) as
was SVO competitive (0.046), t(189) ⫽ 4.79, p ⬍ .01). Scenarios
did not show a significant effect, nor did the other value measures.
A second model featured low-assertiveness social and instrumental
expectancies in Level 1. In this model, instrumental expectancies
were not a significant predictor, but social expectancies were
(⫺0.026), t(189) ⫽ 2.03, p ⬍ .05, as was SVO competitive
(0.054), t(189) ⫽ 3.94, p ⬍ .01. None of the other value measures
were significantly predictive. A third model featured moderate
assertiveness social and instrumental expectancies in Level 1. In
this model, social expectancies were not a significant predictor, but
instrumental expectancies were (⫺0.047), t(189) ⫽ 2.81, p ⬍ .01,
as was SVO competitive (0.046), t(189) ⫽ 5.94, p ⬍ .01. All six
measures of expectancies (in Level 1) were used in a final model
as well as the individual difference measures (in Level 2). SVO
competitive was predictive (0.046), t(184) ⫽ 5.14, p ⬍ .01, as
were low-assertiveness social expectancies (⫺0.023), t(184) ⫽
2.21, p ⬍ .05; medium assertiveness instrumental expectancies
(⫺0.042), t(184) ⫽ 3.12, p ⬍ .01; high assertiveness social expectancies (0.040), t(184) ⫽ 2.89, p ⬍ .01; and high assertiveness
instrumental expectancies (0.064), t(184) ⫽ 5.77, p ⬍ .01.
In summary, across these models, both social and instrumental
expectancies for high assertiveness and SVO competitive were
strongly predictive of preferred assertiveness. The multilevel analyses also showed significant effects for low-assertiveness social
expectancies and moderate assertiveness instrumental expectancies. To explore the possible interaction of values and expectancies, additional multilevel models were run to test for interactions
between the expectancy measures and the social value measures
(care about results, important to get along) as well as unmitigated
communion. No significant interactions emerged.
A final exploratory analysis was conducted, starting with the
dependent variable (averaging preferred assertiveness across the
three scenarios) and identifying individuals in the lowest and
highest thirds of assertiveness. Contrasting the expectancies for
these two groups showed similar social and instrumental expectancies for low and medium assertiveness responses. However, the
groups differed markedly in their expectancies for high assertiveness behaviors where highly assertive individuals expected significantly more positive outcomes.
Discussion
Study 2 built on Study 1 by having participants review a series
of concrete scenarios involving interpersonal conflict. Participants
indicated their preferred level of assertiveness in response to each
scenario (ranging from low to high assertiveness) and also recorded their expected social and instrumental outcomes. As anticipated, expectancies predicted preferred assertiveness: Those who
were comparatively optimistic about outcomes for high assertive
behaviors were more likely to prefer highly assertive responses.
Whereas an effect only for social expectancies was found in Study 1,
effects for both social and instrumental expectancies in predicting
assertiveness were shown in Study 2. These expectancy–
assertiveness links remained after controlling for value measures,
including SVO competitive, which also predicted behavioral preferences. In short, preferences for responding in an assertive way
were predicted not only by interpersonal motivations but also by
an individual’s expectations for how costly or beneficial a highly
assertive response would be.
Moderate and low assertiveness expectancies were not consistently or strongly predictive of behavioral preferences. The overall
pattern of results, and exploratory analyses of expectancies, reinforced what Study 1 suggested: Expected outcomes for high assertive behaviors seem to be a promising operationalization of
idiosyncratic assertiveness expectancies.
Measuring Expectancies to Predict Assertiveness
Studies 1 and 2 provided initial support for the predicted link
between expectancies and assertiveness. Another motivation behind these studies was to explore ways in which assertiveness
expectancies might be meaningfully and efficiently measured. The
results warrant brief consideration before continuing on to additional tests of the expectancy–assertiveness link.
An advantage of the line drawing paradigm in Study 1 is that it
yielded a complete line or curve that identified a perceived optimal
ASSERTIVENESS EXPECTANCIES
level of assertiveness. Conceptually, this is a compelling way of
reflecting expectancies, yet it is decontextualized in its representation of both behaviors and outcomes. Using more concrete behaviors and outcomes has appeal but presents a challenge in that
detailing a full and nuanced range of increments in assertiveness
(e.g., 10 possible detailed responses in an interpersonal conflict)
could make for a cumbersome measure.
Expected outcomes for highly assertive behaviors seem like a
promising measure of expectancies. In Study 1, person-to-person
variance appeared to be highest in anticipated outcomes for extreme behaviors, especially highly assertive behaviors (see top half
of Figure 2). In terms of predicting self-reported assertiveness,
expected social outcomes for highly assertive behaviors fared best
(see the bottom of Figure 2). In Study 2, with concrete scenarios,
expected social and instrumental outcomes for high assertiveness
were consistently predictive of behavioral preferences, whereas
expected outcomes for low and moderate assertiveness were generally not. I return in this article’s General Discussion section to
the matter of why expectancies for highly assertive behavior might
have comparatively robust predictive power.
In Studies 3 and 4, I attempted to link high assertiveness
expectancies to real-world assertive behavior. I used concrete
scenarios in both studies like those in Study 2 to gauge expectancies, using these expectancies as an independent measure to predict
partner reports of participant assertiveness in a dyadic negotiation
(Study 3) and coworker reports of participant assertiveness in the
workplace (Study 4). Thus, whereas expectancies in particular
contexts were used to predict behavior in those fictional contexts
in Study 2, expectancies in one set of fictional contexts were used
to predict behaviors in somewhat different real-world contexts in
Studies 3 and 4.
Study 3
Building on Studies 1 and 2, Study 3 operationalized expectancies
as anticipated social and instrumental outcomes for high assertiveness
behaviors in conflict scenarios (measured in domains different from
that featured in the negotiation and collected several weeks in advance). It was expected that these expectancies would predict negotiation behavior and outcomes, controlling for the effect of values. In
addition, Study 3 went beyond Study 2 by including additional value
and dispositional measures related to assertiveness, including conflict
styles and relevant personality facets.
Method
Participants. Sixty-four MBA students, with a mean age of
29.1 (SD ⫽ 2.7), took part in Study 3 during an elective course on
negotiation. Twenty-nine participants (45.%) were women. Fortyone participants (64.1%) identified themselves as Caucasian, 15
(23.4%) identified themselves as Asian or Asian American, 3
(4.7%) identified themselves as Latino or Hispanic, and 1 (1.6%)
identified him- or herself as African American.
Expectancy measures. Prior to the beginning of the course,
participants indicated their expectancies for two scenarios (two of
the three featured in Study 2). One featured responses to a belowmarket-rate salary offer; the other featured responses to managing
a freelance employee during a severe project deadline. The scenarios each featured a common beginning and three behavioral
conclusions varying in assertiveness, as shown in the Appendix.
1549
For each behavior (low, medium, and high assertiveness), participants indicated expected social outcomes on an 11-point scale
ranging from 1 (strong dislike, distrust) to 11 (strong like, trust).
For the salary negotiation, the item read, “How much do you think
the manager would like, trust, and want to interact with you in the
future?” For the freelancer scenario, the item read, “How much do
you think the freelancer would like, trust, and want to interact with
you in the future?”
Participants also indicated expected instrumental outcomes on
an 11-point scale ranging from 1 (extremely poorly) to 11 (extremely well). For the salary negotiation, the item read, “How well
do you think you would do for yourself in terms of final negotiated
salary?” For the freelancer scenario, the item read, “How well do
you think you would do for yourself in terms of project outcomes
(completed profiles, budget, time, etc.)?”
Conflict measures. A week after the expectancy measure, participants completed a variety of other individual-difference measures. Two of these were focused on the domain of conflict: SVO
(as measured in Study 2) and the Thomas-Kilmann Conflict Mode
Instrument, or TKI. The TKI gauges a person’s style for resolving
conflict along five orientations: accommodating, avoiding, collaborating, competing, and compromising (Thomas & Kilmann,
2002). These orientations are located by crossing two dimensions,
similar to those in the SVO: an individual’s emphasis on satisfying
her or his own concerns, and an individual’s emphasis on satisfying the concerns of the other party. The TKI contains 30 paired
statements that describe two of the five conflict orientation modes.
For each pair, respondents were asked to select the statement that
more aptly described them. For instance, one pair asks participants
to choose between “I propose a middle ground” (Compromising)
and “I press to get my points made” (Competing). Possible scores
on each of the five conflict orientations ranged from 0 to 12.
Other individual-difference measures. Along with the conflict
measures, participants completed several other value and personality measures that might account for assertive behavior. These
included a number of eight-item subscales of the Revised NEO
Personality Inventory (Costa & McCrae, 1992), rated on 5-point
scales ranging from 1 (disagree strongly) to 5 (agree strongly).
Several relevant facets of Extraversion were measured, including
Assertiveness (e.g., “I am dominant, forceful, assertive”) and
Warmth (e.g., “I’m known as a warm and friendly person”).
Several relevant facets of Agreeableness were also measured, including Trust (e.g., “My first reaction is to trust people”), Compliance
(e.g., “I hesitate to express my anger even when it’s justified”), and
Altruism (e.g., “I go out of my way to help others if I can”). In
addition, Angry Hostility, a facet of Neuroticism, was measured (e.g.,
“I often get angry at the way people treat me”). In addition, participants completed measures of unmitigated communion (five items)
and social outcome values (two items), as in Study 2.
Negotiation measures. Two weeks after the expectancy assessment, participants were randomly paired and completed a
two-person face-to-face role play negotiation. One participant in
each pair played the role of a company selling a factory; the other
participant played the role of a buyer. Both participants received
the same general information about the case, including the points
that the seller purchased the plant for $15 million several years
ago, that a recent appraisal of the factory indicated a value of $19
million, and that a similar but new plant recently sold for $26
million. The buyer received confidential information indicating
AMES
1550
that their best alternative to a negotiated agreement with their
partner (also known as a BATNA) was $25 million for building a
new plant. The seller received confidential information indicating
that their BATNA was $17 million for stripping and scrapping the
plant. Thus, the buyer should not have paid more than $25 million,
and the seller should not have accepted less than $17 million. No
other issues were introduced or allowed, making this a single-issue
distributive (i.e., zero-sum) negotiation focused on sale price. All
pairs reached a settlement.
After the negotiation (usually lasting 20 –30 min) was finished,
participants completed ratings of their partner. An initial item
gauged perceived assertiveness (“How assertive was your partner
during the negotiation?”) on a 12-point scale ranging from 1 (not
assertive at all) to 12 (extremely assertive). Next, participants
rated whether they saw their partner as under- versus overassertive
on a 5-point scale, including 1 (very underassertive), 2 (somewhat
underassertive), 3 (appropriately assertive), 4 (somewhat overassertive), and 5 (very overassertive).
Participants subsequently rated their partner’s behaviors on a
6-point scale ranging from 1 (not at all/never) to 6 (a great
deal/always). Items included “They made extreme or bold offers,”
“They made meaningful concessions,” “They displayed competitive, aggressive behavior,” “They took steps to avoid and/or diffuse any tension,” and “They revealed helpful or important information.” Finally, participants rated their trust for their partner (“Do
you feel like you could trust your partner in future negotiations?”)
and their attitudes about future interaction with their partner (“Do
you look forward to future interactions and teamwork with your
partner?”) on 12-point scales ranging from 1 (strongly disagree) to
12 (strongly agree).
Results
Means and reliabilities for the measures are shown in Table 3.
The two social outcome value items were only modestly correlated, r(64) ⫽ .31, p ⫽ .01, and are thus analyzed separately.
Table 3
Scale Means and Reliabilities in Study 3
Construct
Items
M
SD
␣
SVO Prosocial
SVO Individualist
SVO Competitive
TKI Collaborate
TKI Compete
TKI Accommodate
TKI Avoid
TKI Compromise
Unmitigated communion
NEO Assertiveness
NEO Warmth
NEO Angry Hostility
NEO Trust
NEO Compliance
NEO Altruism
—
—
—
—
—
—
—
—
5
8
8
8
8
8
8
3.41
4.72
0.87
5.25
5.36
5.86
6.03
7.42
3.32
3.29
4.05
2.70
3.46
3.05
3.99
3.81
3.95
2.50
2.02
3.22
2.30
3.00
2.17
0.64
0.74
0.66
0.87
0.84
0.57
0.52
—
—
—
—
—
—
—
—
.64
.81
.81
.84
.89
.60
.70
Note. Number of items and alpha values are shown only for Likert-rating
scales. SVO ⫽ Social Value Orientation; TKI ⫽ Thomas-Kilmann Conflict
Mode Instrument; dashes represent scales where the number of items is not
necessarily meaningful.
Expectancies. The independent variable of expectancies was
based on responses to the conflict scenarios. Across a wide range
of results and consistent with Study 2, expected outcomes for the
high assertiveness behaviors (averaged across the two scenarios)
were substantially and consistently more predictive than those for
the low or medium assertiveness behaviors.2 As such, the results
focus on these measures as the operationalization of expectancies.
For social outcomes at high assertiveness, mean expectancies on
the 11-point scale were 5.20 (SD ⫽ 1.66); for instrumental outcomes at high assertiveness, mean expectancies were 7.09 (SD ⫽
1.90).
Predicting assertiveness. The pattern of results was similar
across roles in the negotiation; accordingly, the following results
collapse across role. Partner ratings were used as measures of
behavior (e.g., making concessions). A measure of “deal term”
was created by taking the z score of the settlement price across all
participants and then multiplying this term by ⫺1 for buyers (i.e.,
higher for sellers who sold for comparatively high prices and
higher for buyers who bought for comparatively low prices).
As shown in Table 4, social expectancies were significantly
related to deal terms, but instrumental expectancies were not, nor
were any other measures. In contrast to this pattern, instrumental
expectancies were significantly related to several ratings made by
partners (including target assertiveness, overassertiveness, and
competitive behavior, as well as partner trust for target and partner
interaction attitudes), but social expectancies were not.
Expectancies were correlated with several of the other measures
(see Table 4). Both social and instrumental expectancies were
positively correlated with the TKI competitive orientation and
negatively correlated with the TKI avoiding orientation. Instrumental expectancies, but not social expectancies, were correlated
with TKI accommodating and NEO Assertiveness and Warmth.
Expectancies were not correlated with unmitigated communion or
social value items (care about results, important to get along).
A number of the nonexpectancy measures were correlated with
partner ratings. TKI competitive orientation was positively correlated with partner ratings of participant competitive behavior and
negatively correlated with partner trust and interaction attitudes.
TKI avoiding was negatively correlated with partner ratings of
participant assertiveness and positively correlated with partner
interaction attitudes. NEO Assertiveness was positively related
with partner ratings of participant assertiveness and negatively
correlated with partner trust and interaction attitudes. NEO Angry
Hostility was negatively correlated with interaction attitudes and
2
In general, expectancies for low and medium assertiveness were not
predictive. For instance, social expectancies for low and medium assertiveness were not significantly correlated with concession behavior,
r(59) ⫽ ⫺.03, p ⫽ .82; and, r(59) ⫽ ⫺.04, p ⫽ .78, respectively, or with
deal terms, r(59) ⫽ ⫺.01, p ⫽ .93; and, r(59) ⫽ .06, p ⫽ .63, respectively.
Likewise, instrumental expectancies for low assertiveness were not significantly correlated with concession behavior, r(59) ⫽ .00, p ⫽ .98; and,
r(59) ⫽ ⫺.25, p ⫽ .78, respectively, though instrumental expectancies for
medium assertiveness were directionally correlated with concession behavior, r(59) ⫽ ⫺.25, p ⫽ .06. Instrumental expectancies for low and medium
assertiveness were not significantly correlated with deal terms, r(59) ⫽
⫺.01, p ⫽ .97; and, r(59) ⫽ ⫺.07, p ⫽ .56, respectively. As shown in
Table 4, high assertiveness expectancies were more predictive.
.23†
.06
.10
⫺.05
⫺.26ⴱ
⫺.20
.01
⫺.23†
⫺.11
NEO measures
⫺.07
⫺.10
⫺.12
.01
⫺.09
⫺.17
Values
⫺.09
⫺.01
⫺.21
ⴱ
Note. Expected outcomes are for high assertiveness behaviors. n ⫽ 58 – 60, depending on missing values.
p ⬍ .05. ⴱⴱ p ⬍ .01. † p ⬍ .10.
⫺.14
⫺.09
.14
.04
⫺.07
⫺.11
⫺.05
⫺.03
⫺.13
⫺.03
⫺.11
.17
.21
.13
.18
.12
⫺.22†
⫺.01
.29ⴱ
.12
.20
⫺.03
⫺.19
.08
.23†
.21†
.01
⫺.02
⫺.14
⫺.06
.08
.00
⫺.09
.09
⫺.14
⫺.10
Care about results
Important to get along
Unmitigated communion
⫺.09
⫺.10
.17
.07
⫺.02
.00
.37ⴱⴱ
⫺.12
⫺.19
⫺.11
Conflict styles
⫺.01
⫺.03
⫺.01
.12
⫺.13
.04
⫺.01
⫺.11
⫺.08
⫺.13
.14
⫺.05
⫺.03
⫺.09
.01
⫺.06
.07
Social value orientations
.11
.11
.15
⫺.01
.01
⫺.08
⫺.14
⫺.17
⫺.09
.15
.20
.08
⫺.21
⫺.23†
Assertiveness
Warmth
Angry Hostility
Trust
Compliance
Altruism
⫺.23†
⫺.31ⴱ
.07
.25ⴱ
.25†
.29ⴱ
Expectancies
⫺.05
⫺.09
Concessions
Competitive
behavior
Extreme
offers
Over
assertiveness
.10
.20
.09
⫺.39ⴱⴱ
.03
.11
.27ⴱ
⫺.22†
⫺.31ⴱ
.04
Collaborate
Compete
Accommodate
Avoid
Compromise
.01
⫺.04
.05
.20
.33ⴱ
.40ⴱⴱ
—
⫺.17
.08
.14
Assertiveness
Instrumental
expectancy
.09
.39ⴱⴱ
⫺.28ⴱ
⫺.30ⴱ
⫺.01
⫺.03
.06
⫺.06
—
.40ⴱⴱ
Prosocial
Individualist
Competitive
Social
Instrumental
Variable
Social
expectancy
Table 4
Correlations Between Individual-Difference Measures and Negotiation Behavior and Outcomes in Study 3
.03
⫺.02
.01
.10
⫺.05
⫺.08
.05
⫺.01
⫺.05
⫺.06
⫺.08
.03
⫺.09
.24†
.11
⫺.15
.07
.23†
⫺.25ⴱ
Avoid
tension
⫺.07
.00
.17
.01
.13
.13
.16
.03
.11
.05
⫺.07
.00
.07
.03
.14
⫺.24†
.16
⫺.24†
⫺.19
Shared
helpful info.
.01
.19
⫺.06
⫺.30ⴱ
⫺.09
⫺.18
.12
.14
.01
⫺.12
⫺.27ⴱ
.13
.13
.17
⫺.03
.17
⫺.21
⫺.12
⫺.39ⴱⴱ
Partner
trust
⫺.22†
.22†
⫺.05
⫺.40ⴱ
⫺.04
⫺.28ⴱ
.17
.25†
.10
⫺.23†
⫺.28ⴱ
.06
.27ⴱ
.19
⫺.03
.12
⫺.13
⫺.17
⫺.35ⴱⴱ
Partner
interaction
attitudes
⫺.05
⫺.03
⫺.02
.11
.05
⫺.03
.03
⫺.14
⫺.10
.00
.08
⫺.14
⫺.04
.06
.01
.01
⫺.02
.31ⴱ
.23†
Deal
term
ASSERTIVENESS EXPECTANCIES
1551
AMES
1552
NEO Compliance was negatively correlated with partner ratings of
participant competitive behavior.
Independent effect prediction. To test the independent effect
prediction—that expectancies would remain predictive after controlling for the effects of values—regression models were run in
cases in which both expectancy and value measures were predictive or were correlated with one another. In a model predicting
partner ratings of participant assertiveness with the instrumental
expectancy measure and the TKI constructs for competing, accommodating, and avoiding (i.e., the TKI measures that were correlated with instrumental expectancy), instrumental expectancies
remained predictive of assertiveness (␤ ⫽ .27), t(54) ⫽ 2.07, p ⬍
.05. Of the TKI measures, avoiding remained predictive (␤ ⫽
⫺.33), t(54) ⫽ 2.13, p ⬍ .05, whereas the others did not ( ps ⬎
.18). In a combined model predicting partner ratings of participant
assertiveness, instrumental expectancies remained predictive (␤ ⫽
.28), t(56) ⫽ 2.23, p ⬍ .05, whereas NEO Assertiveness was only
marginally so (␤ ⫽ .23), t(56) ⫽ 1.79, p ⫽ .08.
Similar models were run predicting partner ratings of target
competitive behavior, partner trust, and partner interaction attitudes. Across these models, after controlling for NEO and TKI
measures, as well as SVOs and value measures, instrumental
expectancies remained predictive. In summary, the independent
effect prediction was generally supported. In addition, a wide
range of models were analyzed to test for possible interactions
between various value measures and expectancies (e.g., the interaction of social expectancies and unmitigated communion). No
significant interactions emerged.
Discussion
Study 3 supported the behavioral prediction by linking expectancies to actual interpersonal behavior and outcomes. In a faceto-face zero-sum role play negotiation, those who expected more
positive social outcomes at high levels of assertiveness achieved
better deal terms. Those who expected more positive instrumental
outcomes at high levels of assertiveness were seen as behaving
more assertively. It is worth noting that the measures of participants’ assertiveness were not self-reports (as in Studies 1 and 2)
but rather were appraisals provided by randomly assigned negotiation partners.
Consistent with the independent effect prediction, these effects
of expectancies generally remained after controlling for relevant
value constructs. Although negotiation deal term was predicted by
social expectancies, it was not significantly related to any of the
other constructs (including SVO, conflict resolution style, unmitigated communion, and various NEO personality measures).
Study 4
In Study 4, I sought additional evidence for the link between
assertiveness expectancies and real-world behavior. One important
real-world instantiation of assertiveness is the inauguration of
potentially productive negotiations. As Babcock, Gelfand, Small,
and Stayn (2006) recently argued, the “propensity to initiate negotiations” varies from person to person, with some people habitually bypassing opportunities to negotiate when doing so would be
natural and productive, whereas others chronically inaugurate negotiations, perhaps even when they are inappropriate. In Study 4,
I used a measure of propensity to initiate negotiations that included
constructs that have been linked with real-world negotiation behavior, such as the number of days since a person’s most recent
workplace negotiation as well as days until their anticipated next
workplace negotiation (Babcock et al., 2006). It was predicted that
expectancies for high assertiveness behaviors would be positively
associated with propensity to initiate negotiations measures for
recognition and entitlement and negatively associated with the
measure of apprehension.
Study 4 also gauged real-world assertiveness in the workplace,
using work colleague ratings of targets’ assertiveness, including
their tendency to speak up and to stand their ground in a conflict.
Each participant was rated by multiple colleagues who had typically worked with the participant for several years (participants
were typically 28 –32 years old). This dependent variable is thus a
meaningful measure of real-world assertiveness, parallel to measures that have been linked in recent research on perceptions of
influence and leadership in the workplace (Ames & Flynn, 2007).
In summary, it was expected that assertiveness expectancies
(measured through responses to conflict scenarios, as in Study 3)
would be predictive of the propensity to initiate negotiations and
colleague-rated assertiveness measures and that these links would
remain after controlling for relevant motivations.
Method
Participants. Seventy-five MBA students, with a mean age of
29.0 (SD ⫽ 2.6), took part in Study 4 during a management
elective course. Thirty participants (40.0%) were women. Fortyeight participants (64.0%) identified themselves as Caucasian, 18
(24.0%) identified themselves as Asian or Asian American, 4
(5.3%) identified themselves as Latino or Hispanic, and 1 (1.3%)
identified him- or herself as African American.
Expectancy and value measures. As part of a course exercise,
expectancies were measured as in Study 3, as were social outcome
values, unmitigated communion, and SVOs.
Propensity to initiate negotiations. A separate exercise from
the above gauged the propensity to initiate negotiations using a
measure including three dimensions associated with identifying
and starting negotiations (Babcock et al., 2006). Recognition measures an individual’s tendency to recognize situations in which
negotiation is possible (four items; e.g., “There are many things
available to people, if only people asked for them”). Entitlement
measures the extent to which an individual feels owed or entitled
to receiving an outcome better than the status quo (three items;
e.g., “I think situations should be changed to fit my desires”).
Apprehension measures an individual’s feelings of unrest upon
initiating a negotiation (five items; e.g., “I feel anxious when I
have to ask for something I want”). Items were rated on a 7-point
scale ranging from 1 (strongly disagree) to 7 (strongly agree).
Workplace ratings. As part of participants’ course work in the
prior academic year, they solicited feedback on their behaviors
from former work colleagues (individuals who had worked with
the students before they began their 2-year graduate program). On
average, students recruited four former colleagues (who had typically worked with the student for 2–5 years and indicated they
knew the student well) who provided anonymous feedback, rating
various behaviors on a 7-point scale ranging from 1 (never) to 7
(always). Assertiveness in the workplace was gauged with three
ASSERTIVENESS EXPECTANCIES
items that reflected everyday usage of the term (see Ames &
Flynn, 2007), “S/he speaks up and shares his/her views when it is
appropriate,” “S/he is able to stand his/her ground in a heated
conflict,” and “S/he is willing to engage in constructive interpersonal confrontations.”
Results
Scale means and reliabilities are shown in Table 5. The two
social outcome value items were only moderately correlated,
r(74) ⫽ .32, p ⬍ .01, and, as in Studies 2 and 3, are reported
separately. As in Study 3, expectancies for the high assertiveness
behaviors were substantially and consistently more predictive than
those for the low or medium assertiveness behaviors.3 As such, the
results focus on these measures (averaged across the two scenarios) as the operationalization of expectancies. For social outcomes
at high assertiveness, mean expectancies on the 11-point scale
were 5.30 (SD ⫽ 1.70); for instrumental outcomes at high assertiveness, mean expectancies were 7.22 (SD ⫽ 1.84).
Predicting propensity to initiate negotiations. Expectancies
were predicted to correlate with the propensity to initiate negotiations constructs. As shown in Table 6, this was significant only in
the case of entitlement. Both social and instrumental expectancies
were positively correlated with entitlement. Contrary to predictions, expectancies were not significantly correlated with recognition or apprehension.
The SVO measure of prosocial orientation was also correlated
with entitlement. In a model predicting entitlement with instrumental expectancies and SVO prosocial, expectancies remained
significantly predictive (instrumental expectancy ␤ ⫽ .38), t(71) ⫽
3.40, p ⬍ .01, whereas SVO prosocial was directionally predictive
(␤ ⫽ ⫺.20), t(71) ⫽ ⫺1.79, p ⫽ .08. In a model predicting
entitlement with social expectancies and SVO prosocial, expectancies were significantly predictive (social expectancy ␤ ⫽ .27),
t(71) ⫽ 2.37, p ⬍ .05, as was SVO prosocial (␤ ⫽ ⫺.27), t(71) ⫽
2.34, p ⬍ .05.
Predicting workplace assertiveness. Social expectancies were
significantly correlated with workplace assertiveness (see Table 6).
The more people expected positive social outcomes for highly assertive behaviors in a series of fictional scenarios, the more their former
work colleagues saw them as assertive in the workplace, consistent
with the behavioral prediction. Instrumental expectancies were not
significantly correlated with workplace assertiveness.
Table 5
Scale Means and Reliabilities in Study 4
Construct
Items
M
SD
␣
Unmitigated communion
SVO Prosocial
SVO Individualist
SVO Competitive
Apprehension
Recognition
Entitlement
Work assertiveness
5
—
—
—
5
4
3
3
3.33
3.39
4.83
0.79
3.96
5.63
3.67
5.64
0.66
3.79
3.91
2.35
1.57
0.88
1.09
0.56
.64
—
—
—
.93
.72
.57
.85
Note. SVO ⫽ Social Value Orientation; dashes represent scales where
the number of items is not necessarily meaningful.
1553
The value measures were generally unrelated to coworkers’ ratings
of participants’ assertiveness in the workplace (see Table 6). Apprehension was negatively correlated with assertiveness. Although
not necessarily a value measure, it is worth examining whether the
apprehension measure usurped the predictive power of expectancies. A regression model was run, featuring both social expectancies and apprehension predicting workplace assertiveness. Both
expectancies (␤ ⫽ .33), t(67) ⫽ 2.89, p ⬍ .01, and apprehension
(␤ ⫽ ⫺.21), t(67) ⫽ ⫺1.87, p ⫽ .07, were predictive in the
expected directions. These results are consistent with the independent effect prediction. A wide range of models were analyzed to
test for possible interactions between various value measures and
expectancies. As in Study 3, no significant interactions emerged.
Discussion
In Study 4, work colleagues of participants reported their perceptions of the participants’ assertiveness in the workplace. This
measure of real-world assertiveness was predicted by targets’
social expectancies on the basis of responses to scenarios featuring
interpersonal conflicts. The separation in time and context of these
measures bears reiteration: Work colleagues provided assertiveness ratings of targets typically based on several years’ worth of
acquaintanceship in a work context, and this measure was predicted by targets’ self-reported expectancies on the basis of fictional conflict scenarios over 1 year after having left the workplace. This expectancy– behavior link remained after controlling
for motives, including unmitigated communion and SVOs.
Participants also reported their propensity to initiate negotiations, components of which have been linked to real-world negotiation behavior. The measure of entitlement showed a link to
expectancies, though measures of recognition and apprehension
did not. The link between entitlement and expectancies remained
after controlling for other measures, including SVO.
General Discussion
Why do some people come on strong in interpersonal conflicts,
whereas others more readily capitulate? The traditional answer in
much of the literature on conflict and negotiation revolves around
interpersonal and social values. In this article, I have attempted to
broaden the answer by invoking another construct psychologists
have long turned to in explanations of behavior: expectancies. Four
studies shed light on the form of assertiveness expectancies and on
their links with behavior. In Study 1, a line drawing paradigm was
3
As in Study 3, expectancies for low and medium assertiveness were
generally not as predictive as expectancies for high assertiveness behaviors. For instance, social expectancies for low and medium assertiveness
were not significantly correlated with entitlement, r(74) ⫽ ⫺.08, p ⫽ .51;
and, r(74) ⫽ ⫺.12, p ⫽ .32, respectively, or with workplace assertiveness,
r(74) ⫽ .19, p ⫽ .12; and, r(74) ⫽ .08, p ⫽ .51, respectively. Instrumental
expectancies for medium assertiveness were not significantly correlated
with entitlement, r(74) ⫽ ⫺.01, p ⫽ .96, though instrumental expectancies
for low assertiveness were directionally correlated with entitlement,
r(74) ⫽ ⫺.19, p ⫽ .10. Instrumental expectancies for low and medium
assertiveness were not significantly correlated with workplace assertiveness, r(74) ⫽ .09, p ⫽ .48; and, r(74) ⫽ ⫺.07, p ⫽ .57, respectively. As
shown in Table 6, high assertiveness expectancies were predictive.
AMES
1554
Table 6
Correlations Between Expectancies, Values, Propensity to Initiate Negotiations, and Workplace Assertiveness in Study 4
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Social expectancies
Instrumental expectancies
Care about results
Important to get along
Unmitigated communion
SVO Prosocial
SVO Individualist
SVO Competitive
Recognition
Entitlement
Apprehension
Work assertiveness
1
2
3
4
5
6
7
8
9
10
11
12
—
.38ⴱⴱ
⫺.05
.02
⫺.15
.05
⫺.01
⫺.08
.21†
.23ⴱ
⫺.18
.37ⴱⴱ
—
.03
.08
⫺.09
⫺.14
.06
.13
.18
.39ⴱⴱ
⫺.20†
.11
—
⫺.32ⴱⴱ
⫺.18
⫺.38ⴱⴱ
.31ⴱⴱ
.09
.18
.00
⫺.12
⫺.06
—
.03
.20†
⫺.15
.07
⫺.31ⴱⴱ
.06
.25ⴱ
⫺.19
—
.23†
⫺.28ⴱ
.09
⫺.13
⫺.13
.11
⫺.19
—
⫺.81ⴱⴱ
⫺.26ⴱ
⫺.14
⫺.25ⴱ
⫺.14
.03
—
⫺.35ⴱⴱ
.07
.23†
.18
.00
—
.11
.02
⫺.07
⫺.05
—
.27ⴱ
⫺.42ⴱⴱ
.09
—
.04
.04
—
⫺.28ⴱ
—
Note. n ⫽ 74 except for correlations with Work assertiveness, where n ⫽ 70 because of missing cases. SVO ⫽ Social Value Orientation.
p ⬍ .05. ⴱⴱ p ⬍ .01. † p ⬍ .10.
ⴱ
used and revealed that most participants expected a curvilinear,
inverted-U-shaped, link between assertiveness and outcomes. That is,
they thought they could push up to point, but no further without
incurring costs. This effect was consistent across both social and
instrumental domains and for both undergraduate student and
MBA student populations. This curvilinear pattern also echoes
evidence of the actual impact of interpersonal assertiveness provided by Ames and Flynn (2007). However, although many participants displayed the same overall curved shape for their expectancies, considerable individual variance emerged in the parameters
of the expectancies, such as the perceived optimal level of assertiveness. A link between the variance in social expectancies and
self-reported assertiveness was shown in Study 1. More important,
variance in expectancies was not closely related to variance in
values (e.g., concern with relationships), and the link between
expectancies and self-reported assertiveness remained after controlling for values.
This pattern extended across the remaining studies, which
showed that expectancies had a distinct ability to predict assertive
choices in a series of conflict scenarios (Study 2) and attitudes
related to initiating real-world negotiations (Study 4). Measures of
actual behaviors likewise showed the predicted effects: Expectancies predicted partner ratings of behavior, as well as deal terms, in
a face-to-face dyadic negotiation (Study 3). Expectancies also
predicted work colleague ratings of participants’ assertiveness in
the workplace (Study 4). Both the negotiation and workplace
studies featured links between targets’ self-reported expectancies
and other people’s perceptions of the targets’ assertiveness. In
short, across multiple samples and multiple measures of assertiveness, expectancies played a noteworthy predictive role that was
distinct from values, including measures such as SVOs, unmitigated communion, and conflict styles. The present results suggest
that expectancies can help researchers expand their answer to the
question of why some people push harder than others.
None of this is to imply, of course, that values or motivations do
not matter or do not help to explain assertiveness. On the contrary,
the analyses reported here tended to highlight simultaneous effects
for both expectancies and values. Hybrid models may best predict
assertiveness by capitalizing on what social cognitive theories of
personality have shown: Behavior flows from perceivers taking
action on their goals in light of their idiosyncratic perceptions of
circumstances (e.g., Mischel & Shoda, 1995).
The present results did not show evidence of interactions between expectancies and values. It stands to reason that expectancies about highly valued outcomes (relational, instrumental, or
other) will predict behavior more than expectancies about unvalued outcomes. The present work may not have revealed these
effects because the social and relational outcomes in question were
at least reasonably valuable to more or less all respondents (in
effect, a restriction of range in value). Perhaps future work will
uncover value-expectancy interactions through more focused
measurement or situations featuring wider variance in desired
outcomes.
Lessons for Studying Assertiveness Expectancies
The present results are hardly the final word on assertiveness
expectancies, though they seem to suggest some directions for
researchers on several different topics, such as how to measure
expectancies effectively . The studies reported here revealed the
potential for operationalizing assertiveness expectancies through
expected outcomes for highly assertive behaviors. But would expectancies for even more assertive behavior be even more predictive? Probably not. The high assertive behaviors featured in this
article (e.g., in Study 2) are forceful, but not so extreme or
pathological that the participant population overwhelmingly rejected them as outrageous. At noticeably higher levels of assertion
(e.g., outrageous demands, belligerence), it is likely that perceivers
would increasingly share negative expectancies, and thus such
expectancies would likely not be diagnostic of perceived optimal
assertiveness or predictive of assertive behavior. For high assertiveness expectancies to be predictive, the behaviors they reflect
need to be assertive enough to repel some individuals but not so
assertive that they repulse virtually everyone.
Why would high assertiveness expectancies be especially predictive? A basic answer is that between-person variance may be
greater for high assertiveness expectancies than low assertiveness
ones. People may generally agree that low assertiveness approaches often fail and that moderate assertive approaches are
reasonably effective; they may disagree most strikingly about what
ASSERTIVENESS EXPECTANCIES
happens when one pushes hard. Yet, why would they disagree? It
could be that aggressive and accommodating people have
markedly different evidence about the outcomes of high assertiveness. Unassertive individuals might avoid assertive behavior and
thus never receive clear evidence about the consequences that
follow, being left instead to “catastrophize” what would have
happened had they pushed harder (e.g., Vasey & Borkovec, 1992).
Assertive individuals, howwever, might engage in aggression but
selectively interpret the apparent subsequent evidence, convincing
themselves that their action was effective or even necessary. This
asymmetry might not exist at the other end of the behavioral
spectrum: Both low- and high-assertive individuals may find the
disappointing or dismal outcomes obvious when they capitulate or
yield. Thus, expectancies and experience could reinforce one another at high levels of assertiveness, more so than at low levels,
leading more and less assertive individuals to diverge and sustain
potentially exaggerated perceptions.
Both social and instrumental expectancies were gauged in all four
of the studies reported here. Although social and instrumental measures tended to be moderately correlated with each other, they showed
somewhat different abilities to predict other measures. In Study 1,
social expectancies predicted self-rated assertiveness, though instrumental expectancies did not. In Study 2, both social and instrumental
expectancies were predictive of behavioral preferences. In Study 3,
social expectancies were predictive of deal terms, whereas instrumental expectancies were predictive of behaviors and partner perceptions.
In Study 4, social expectancies predicted work assertiveness, whereas
instrumental expectancies did not. On balance, one might conclude
that social expectancies performed slightly better as a predictor. There
may be psychological reasons for this (e.g., most people are generally
very sensitive to social outcomes, or at least their perceptions of them)
as well as measurement reasons (e.g., how these constructs were
operationalized in the present studies). In order to learn more about
the nature and impact of expectancies, it seems important to measure
them in both the social and instrumental domains. Researchers might
also attend to individual differences that moderate which expectancy
is more or less predictive of behavior for different individuals.
The ability of expectancies to predict behavior naturally raises
another question: What predicts—that is, generates or shapes—
expectancies? It seems likely that evidence and feedback play
some role in updating expectancies. Expectancies also likely have
some deeper core tracing back to basic working models for the
self, others, and relationships that are built up over the course of a
lifetime (e.g., Campbell, Simpson, Boldry, & Kashy, 2005). Baldwin and Keelan (1999) argued that individuals higher in trait
self-esteem had more positive interpersonal expectancies about
their own ability to secure affiliation from others. An exploratory
analysis of the present results echoes this, showing some evidence
that those lower in self-esteem were more pessimistic about their
ability to pursue their interests without suffering relational costs.4
Future research could further examine these connections with
working models for the self and others, revealing more about
where assertiveness expectancies come from and, as a result, how
they operate and evolve.
Conclusion
The present work brings together two vigorous and noteworthy
traditions of research: one committed to explaining why people act
1555
with a given level of assertiveness in conflicts and negotiations and
another invoking interpersonal expectancies as predictors of behavior
and relations. The combination has substantial promise. The expectancies examined here showed a meaningful ability to predict assertive
behavior in interpersonal conflict, helping to round out researchers’
understanding of who pushes hard or holds back, and why. These
expectations arguably deserve a place alongside motivations in psychologist’s models and, arguably, our interventions.
4
Although not a focus of the present research, self-esteem was measured
in a subsample of Study 1: The MBA participants completed the Single
Item Self-Esteem measure (Robins, Hendin, & Trzesniewski, 2001). This
was correlated with the social optimum measure of assertiveness at
r(101) ⫽ .17, p ⫽ .09. The correlation with the instrumental optimum
measure was not significant.
References
Amanatullah, E. T. (2008). Negotiating assertively: Exploring benefits of
representing another to women in salary negotiations. Manuscript in
preparation.
Amanatullah, E. T., Morris, M. W., & Curhan, J. R. (in press). Negotiators
who give too much: Unmitigated communion, relational anxieties, and
economic costs in distributive and integrative bargaining. Journal of
Personality and Social Psychology.
Ames, D. R., & Flynn, F. J. (2007). What breaks a leader: The curvilinear
relation between assertiveness and leadership. Journal of Personality
and Social Psychology, 92, 307–324.
Babcock, L., Gelfand, M., Small, D., & Stayn, H. (2006). Gender differences in the propensity to initiate negotiations. In D. De Cremer, M.
Zeelenberg, & J. K. Murnighan (Eds.), Social psychology and economics
(pp. 239 –259). Mahwah, NJ: Erlbaum.
Baldwin, M. W., & Dandeneau, S. D. (2005). Understanding and modifying the relational schemas underlying insecurity. In M. Baldwin (Ed.),
Interpersonal cognition (pp. 33– 61). New York: Guilford Press.
Baldwin, M. W., & Keelan, J. P. R. (1999). Interpersonal expectations as
a function of self-esteem and sex. Journal of Social and Personal
Relationships, 16, 822– 833.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral
change. Psychological Review, 84, 191–215.
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for
interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529.
Bowles, H. R., Babcock, L., & Lai, L. (2007). Social incentives for gender
differences in the propensity to initiate negotiation: Sometimes it does
hurt to ask. Organizational Behavior and Human Decision Processes,
103, 84 –103.
Campbell, L., Simpson, J. A., Boldry, J. G., & Kashy, D. (2005). Perceptions of conflict and support in romantic relationships: The role of
attachment anxiety. Journal of Personality and Social Psychology, 88,
510 –531.
Carnevale, P. J., & De Dreu, C. K. W. (2006). Motive: The negotiator’s
raison d’être. In L. Thompson (Ed.), Frontiers of social psychology:
Negotiation theory and research (pp. 55–76). New York: Psychology
Press.
Carnevale, P. J., & Pruitt, D. G. (1992). Negotiation and mediation. Annual
Review of Psychology, 43, 531–582.
Costa, P. T., Jr., & McCrae, R. R. (1992). The Revised NEO Personality
Inventory (NEO-PIR) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources.
1556
AMES
De Cremer, D., & Leonardelli, G. J. (2003). Cooperation in social dilemmas and the need to belong: The moderating effect of group size. Group
Dynamics: Theory, Research, and Practice, 7, 168 –174.
De Dreu, C. K., & Van Lange, P. A. (1995). Impact of social value
orientation on negotiator cognition and behavior. Personality and Social
Psychology Bulletin, 21, 1178 –1188.
De Dreu, C. K., Weingart, L. R., & Kwon, S. (2000). Influence of social
motives on integrative negotiation: A meta-analytic review and test of two
theories. Journal of Personality and Social Psychology, 78, 889 –905.
Downey, G., & Feldman, S. I. (1996). Implications of rejection sensitivity
for intimate relationships. Journal of Personality and Social Psychology,
70, 1327–1343.
Fehr, B., Baldwin, M., Collins, L., Patterson, S., & Benditt, R. (1999).
Anger in close relationships: An interpersonal script analysis. Personality and Social Psychology Bulletin, 25, 299 –312.
Fritz, H. L., & Helgeson, V. S. (1998). Distinctions of unmitigated communion from communion: Self-neglect and overinvolvement with others. Journal of Personality and Social Psychology, 75, 121–140.
Kammrath, L. K., & Dweck, C. (2006). Voicing conflict: Preferred conflict
strategies among incremental and entity theorists. Personality and Social
Psychology Bulletin, 32, 1497–1508.
Komorita, S. S., & Parks, C. D. (1996). Social dilemmas. Boulder, CO:
Westview Press.
Locke, K. D., & Sadler, P. (2007). Self-efficacy, values, and complementarity
in dyadic interactions: Integrating interpersonal and social-cognitive theory.
Personality and Social Psychology Bulletin, 33, 94 –109.
McClintock, C. G., & Liebrand, W. B. (1988). Role of interdependence
structure, individual value orientation, and another’s strategy in social
decision making: A transformational analysis. Journal of Personality
and Social Psychology, 55, 396 – 409.
Messick, D. M., & McClintock, C. G. (1968). Motivational basis of choice
in experimental games. Journal of Experimental Social Psychology, 3,
85–101.
Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252–283.
Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of
personality: Reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological Review, 102, 246 –268.
Olekalns, M., & Smith, P. L. (2003). Testing the relationships among
negotiators’ motivational orientations, strategy choices and outcomes.
Journal of Experimental Social Psychology, 39, 101–117.
Raudenbush, S., Bryk, A., Cheong, Y. F., & Congdon, R. (2004). HLM 6:
Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software International.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring
global self-esteem: Construct validation of a single-item measure and the
Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 27, 151–161.
Roese, N. J., & Sherman, J. W. (2007). Expectancy. In A. W. Kruglanski
& E. T. Higgins (Eds.), Social psychology: A handbook of basic principles (Vol. 2, pp. 91–115). New York: Guilford Press.
Thomas, K. W., & Kilmann, R. H. (2002). Thomas-Kilmann Conflict Mode
Instrument. Palo Alto, CA: CPP, Inc.
Tolman, E. C. (1932). Purposive behavior in animals and men. New York:
Appleton-Century-Crofts.
Van Lange, P. A. M. (1999). The pursuit of joint outcomes and equality in
outcomes: An integrative model of social value orientation. Journal of
Personality and Social Psychology, 77, 337–349.
Van Lange, P. A. M., Otten, W., De Bruin, E. M. N., & Joireman, J. A.
(1997). Development of prosocial, individualistic, and competitive orientations: Theory and preliminary evidence, Journal of Personality and
Social Psychology, 73, 733–746.
Vasey, M., & Borkovec, T. D. (1992). A catastrophizing assessment of
worrisome thoughts. Cognitive Therapy & Research, 16, 505–520.
White, J. B., Tynan, R., Galinsky, A. D., & Thompson, L. (2004). Face
threat sensitivity in negotiation: Roadblock to agreement and joint gain.
Organizational Behavior and Human Decision Processes, 94, 102–124.
Appendix
Scenarios and Endings for Studies 2, 3, and 4
Salary Negotiation
Scenario beginning: Imagine that you are negotiating with a potential employer about the terms of their job offer to you. The
manager you’d be working with calls you on the phone with the offer,
including a salary of $92,000. This figure might be acceptable to you
but is noticeably below what you know similar people make in similar
positions ($110,000 –$120,000). The job is your top choice, but you
were expecting a considerably higher offer. You are fairly sure there
is room for negotiation. The manager asks you if you’re ready to
accept.
High assertiveness behavior: You make a counteroffer of $120,000.
Medium assertiveness behavior: You make a counteroffer of
$105,000.
Low assertiveness behavior: You accept the existing $92,000
offer without further negotiation.
Managing Freelancer
Scenario beginning: Imagine that you are working in a consulting firm and are preparing a series of company profiles for an
important client. A freelance graphics artist who you work with
regularly (you’re his major customer, and he’s your most effective
freelancer) has been taking the raw text and figures you’ve been
sending by e-mail and creating the presentation-ready profile documents. Each one takes about an hour, and the freelancer charges
you $80 an hour ($80 per profile). The freelancer sent you what he
thought was the final profile and said he was planning to take a few
much-needed days off. But you’ve just realized there are 10 more
profiles to do. You don’t know what happened, or who, if anyone,
is to blame, but the e-mail you sent to him with this final batch
didn’t get through. You have 24 hours before everything needs to
be ready. It’s possible that your company’s internal graphics artists
could help, but they are often slow and unreliable.
High assertiveness behavior: You thank the freelancer and
explain the situation. You tell him that without additional
evidence, you can’t be sure this wasn’t his mistake. You remind
him that you are his major customer and that you expect his
flexibility and support under deadlines. You tell him that you
need the 10 profiles done in the next day and that you’ll pay the
regular rate ($800).
ASSERTIVENESS EXPECTANCIES
Medium assertiveness behavior: You thank the freelancer for his
work and explain the situation to him. You tell him you’d be
grateful if he could help out in any way. You offer to pay him for
10 hours ($800) if he could do just five of the profiles (normally
5 hours of work).
Low assertiveness behavior: You thank the freelancer for his
work and tell him you hope he has a good break. You begin to
contact your internal graphics department to see if they can help.
Team Meeting
Scenario beginning: Imagine that you’re in a team meeting with
four colleagues on a strategic planning committee in a mid-sized
manufacturing organization. It’s the first meeting of the team, and
you don’t know the other team members that well. The discussion
is revolving around potential new markets to enter, and one of the
team members begins to strongly advocate entering a new country
with your current products. He goes on for a few minutes very
energetically and then concludes, “This is a huge opportunity just
waiting for us. We’d be idiots to not pursue this.” A year ago, you
and a subordinate did a quick analysis of this same country’s
market. Although the market appeared to be sizable and growing
somewhat quickly, you discovered that the regulatory hurdles were
significant and that the competitive response would be considerable. You also saw major challenges in arranging distribution. You
concluded that short-term success would be difficult, and there was
1557
a real potential for wasting considerable resources better invested
elsewhere. Your teammate just finished advocating for the market,
reiterating with intensity, “Seriously, we’d be fools to let this slip
away.”
High assertiveness behavior: You say, “I totally disagree. This
market is a complete dead-end for us and a potentially huge waste
of time and resources. Believe me, I’ve already looked it over.
There are dozens of other better uses for our money. You won’t get
me to sign off on any strategic plan that includes this market. We
should focus this committee’s attention elsewhere.”
Medium assertiveness behavior: You say, “I can see why this
market could look attractive from the outside, especially given its
growth. A while back, I looked into this opportunity and I’d be
happy to share the details of this research with you. But my
recollection is that we identified multiple barriers that made this
opportunity look unattractive. I welcome everyone’s opinion on
this, but I think we should proceed with caution.”
Low assertiveness behavior: You hold back and say nothing,
allowing the conversation to continue, hoping that with further
analysis, the truth about the market will come out and that it won’t
be attributed to you.
Received December 27, 2006
Revision received June 16, 2008
Accepted June 27, 2008 䡲